Network management has higher and higher requirements on precision operation, suppression of un-allowed services, network security guard and service resource management, which has accelerated the development of service identification technology, represented by, in the device layer, integration of a DPI (Deep Packet Inspection) service identifying board in more and more network devices. The so called “deep” in the DPI service identifying card is relative to the message analysis level of the common L2-L4 switcher/router. The common L2-L4 switcher/router analyzes the content below layer 4 in an IF packet including a source address, a destination address, a source port, a destination port and the protocol type, while the DPI service identifying board does not only analyze the preceding layers but also has an additional function of identifying data packet content.
With the growing expansion of network scale and the growing increase of service flux, the demand for upgrading service identifying network to increase the volume of the service identifying network has become more and more urgent. In the existing upgrading scheme, the scheme of adding DPI service identifying boards to the original switching device in the service identifying network may decrease the cost for upgrading the service identifying network, and is accepted by operators. However, this scheme will cause a switching device to have several DPI service identifying boards, so the master control board of the switching device needs to allocate equally the flux needing to be identified to each DPI service identifying board of the switching device to avoid the overload of a certain DPI service identifying board or certain DPI service identifying boards caused by unequal allocation of the service identifying flux. The usual method for allocation is static configuration, which is specifically: configuring each path for accessing the switching device for each DPI service identifying board respectively, and each DPI service identifying board processing only the service flux on the path configured to itself. The method performs configuration based on paths, but the service flux on paths changes all the time, and the allocation of the service flux can not be equal, so the problem of the overload of DPI service identifying boards caused by the unequal allocation of the service identifying flux still can not be avoided efficiently.